September 1971

Institut für Neutronenphysik und Reaktortechnik

Institut für Angewandte Kernphysik

KFK 1532

The Accelerator Booster Research Reactor

- a Useful Tool for Small Laboratories

K. H. Beckurts, H. G. Eckert, M. Küchle,

K. Scharmer, H. Würz

H.

FederalRepublicofGermany

Ministry ofEducation andScience

FOURTH UNITED NATIONSINTERNATIONAL CONFERENCEON THE PEACEFUL USESOF ATOMIC ENERGY

Geneva, Switzerland, 6-16 September 1971

ADDITIONAL PAPER

AED-CONF­71-100-14GERMANYMay 1971

THEACCELERATOR BOOSTER RESEARCH REACTÖR---,

A USEFULTOOL FORSMALLLABORATORIES

K.R. Beckurts, H.G.Eckert, M. Küchle,

K. Scharmer,If. Würz

Without guarantee for the correctness, exactness and completeness ofthe material, nor for private rights of an y third party.

Private prints,

Distributed as report under AED-CONF-71-100

by

Zentralstelle für Atomkernenergie-Dokumentation (ZAED)

Postal address: 7501 Leopoldshafen, Kernforschungszentrum

Printed in West Germany

--------------

AED-Conf.-71-100-14

The Accelerator Booster Research Reactor ­

a Useful Tool for Small Laboratories

K.H. Beckurts+), H.G. Eckert, M. Küchle, K. Scharmer,H. Würz

Kernforschungszentrum KarlsruheFederal Republic of Germany

+) now at Kernforschungsanlage JülichFederal Republic of Germany

1. INTRODUCTION

Investigations of condensed matter by neutron scatteringmeasurementsconstitute a field of research of growing impor­tance which requires thermal neutronsources of increasinglyhigher intensity. With the present generation of high flux re­actors a limit seems to have been reached of what can be justi­fied economically.For even higher fluxes, pulsed neutron sour­ces in combination with time-of-flight experiments are moresuitable because of the substantial reduction in reactor power.Thus, repetitively pulsed reactors will probably be a promisinglineinthe future development of research reactors.

Two versions are possible:

1. The neutron pulse is generateq by modulating the reactivityinto the prompt super-critical domain.

2. The neutron pulse of an accelerator is multiplied in a so­called booster the reactivity of which is modulated to sup­press the background between successive pulses.

However, the successful utilization of this type of reactorsmust be based on mueh broader experience in the correspondingexperimental techniques and in the problems arising from the re­actor plus the experiment.

-

- 1 -

AED-Conf.-71-100-14

This can be achieved with facilities less ambitious and ex­pensive than SORA and IBR-2 /-1 7. In addition, a facility with,e.a., 100 kW can provide a neutron intensity equivalent to a101 4 n/cm2s flux of a steady-state reactor and would allow mea­ningful research contributions. Such a facility can also be in­stalled and operated by small laboratories.

If a LINAC with a neutron intensity of about 2 x 1011 n/sand 100 cycles per second is already available, the most reaso­nable way is to boost the LI~AC by neutron multiplication. Forthis purpose, the accelerator target is placed in the center ofa small fast reactor core whose reactivity is modulated so as tohave high multiplication during neutron injection and low multi­plication between successive pulses to suppress background. Thefast neutrons leaking out from the core are thermalized in hy­drogenous moderators located in the reflector surrounding thecore. Several beam tubes look at the surface of the moderators.Pulse width and intensity strongly depend on moderator size andmaterial, add moderator optimization is very important.

In Section 2, a typical small booster design is brieflydescribed. It is based on a feasibility study recently made atthe Karlsruhe Nuclear Research Center. In Section 3, new experi­mental and theoretical investigations on the optimization ofpromising moderators are reported.

2. STUDY OF A 100 kW ACCELERATOR BOOSTER

A feasibility study of a repetitively pulsed fast re­search reactor with neutron injection has been performed. Anartist's view of the design is shown in Figure 1. In the text,numbers in brackets refer to this figure. The design is basedon a 60 MeV, 100 mA linear electron accelerator with a pulselength of 7 usec and 100 sec-1 repetition rate. The neutronsource strebgth would be 2.5 x 101 6 n/sec and a substantialmultiplication is r~quired to get thermal neutron fluxes sui­table for condensed matter studies. However, multiplication (M)increases the fast pulse width l·M, (where 1 is the fast neu­tron lifetime). Figure of merit considerations show that l·Mmust be smaller than or equal to (=) , C) being the variance of ther­mal neutron emission time distribution. In a practical case,G = 30/usec and 1 = 35 nsec result in M~850, but for safetyreasons M was limited to 500 or y = -0.002 below prompt cri­tical. To achieve a signal-to-background ratio of about 10,reactivity isdecreased to- 10 $ betweenthe pulsesby meansof a rotating reflector part (1) similar to the SORA-design.Since, in this case, reactivity modulation only has to reducethe background, a rather simple and very conservative two­armed rotor can be used:

- diameter

- frequency

- circumferential velocity

- material

- reactivity

- 2 -

128 cm

50 rps

2 x 104 crn/sec

TIAC6V4

0.2 Ä k/k (=10$ for Pu-239)

AED-Conf.-71-100-14

This relatively slow rotor leads to a time dependent de­layed neutron background which peaks when the rotor passes thecore, Thus, the overall signal-to-background ratio becomes about8.5 for plutonium fuel. These characteristics result in aboosterpower of 100 kW.

The hexagonal reactor core (2) is CO2-cooled and fueledwith PU02 pins (3). To keep neutron lifet1me and critical masssmall, the core is designed as compact as possible. It is sur­rounded by 6 cm of tungsten (4). Two parts of this reflectorcan be radially displaced as safety elements (5). Fine controlis performed by two rotating rods (6). The LINAC target (7) islocated in the center of the core. It consists of helium cooledU-238 shot. The electron beam enters the core from below (8). Acold moderator (parahydrogen) (9) and a moderator at room tem­perature(H20 or zirconium hydride) are located close to thecore and viewed by six beam tubes (10). The reactor block isshielded by concrete, which leaves enough space around the coreaccessible from thetop for maintenance and easy replacement ofcomponents. All parts of the reactor can be exchanged with re­latively simple and inexpensive equipment.

Reactor circuits and secondary equipment are located belowthe experimental floor (11), but enough space is left there forexperiments using the LINAC without booster. Thenthe deflectingmagnet (12) is switched off.

Booster Characteristics

Core power

critical mass

active core height

core volume

number of fuel pins

pin diameter

pitch-to-diameter ratio

core coolant

coolant inlet velocity

outlet pressure

hot spötternperature

target power

target coolant

coolant velocity in the target

3. MODERATOR OPTIMIZATION

100 kW

35 kg Pu-239

20 cm

6.8 1

162

1.4 cm

1.05

CO225 rn/sec

5 bar

SOOoC

6 kW

helium

25 rn/sec

The moderator of apulsed reactor must be optimized forhigh thermal flux and short neutron pulsewidth. The optimumvalues depend on the moderating material and the geometry.Thereföre, extensive calculations of the leakage flux and the

- - . - - "'"--3 -~ ...

AED-Conf.-71-100-1~

variance of the neutron emission time distribution were per­formed for various hydrogenous materials. In order to check thecalculations measurements were performed on some moderators incylindrical geometries.

Parahydrogen is öf particular interest_because of itsunique scattering cross section behaviour L 2-1. Once the neu­trons are slowed down below 15 meV, the moderator becom~ verytransparent. This leads to a high leakage intensity and a shortneutron emission time.

3.1 Methods of Calculation

For calculation of the variance of the thermal neutronemission time distribution the first two time moments must beknown. As shown by Profio et al. L-3-1, these can be calculatedby Laplace transform of the time dependent Boltzmann equation.For moderators whose collision time is comparable with the life­time, transport theory must be used. In order to account proper­ly for the time~of-flight of the neutrons from all directions,a two-dimensional version of the THERM05-code in cylindrical geo­metry has been written which solves the following equations:

00

N(O)= J [PN (0)....,

Ndt = T + 5J0

00 rPN (l)

I.

s>lN (1) = J tRdt TIA\'- ~ I (PN(O) += - v,

L ..J0

<D . \ I .:N(2)= J t2Ndt = T [PN(2l_2 1~-lt'1 PN(1} +( I~-l.( I) 2 (PN(o)

v v.J

0

with the transport and collision operators, respectively,, I i I \

TN = j du-T (L(,J..r,v)N (46", v) ; PN = J dv'P(v,v' ,\.\} N(l(",v').

Leakage fluxes perpendicular to the moderator surfaceand normalized to reactor power were calculated in realisticbooster geometry with the ~wo-dimensional multigroup SN-code5N~W in 58 approximation L 4_/.

Thermal and epithermal group constants were obtained fromTHERM05 calculations. For parahydrogen the Young-Koppel scat­tering kernel was used L-5_7.3.2 Measurements

Time dependent spectra were measured by the time-of­flight method with a cornbination of pulsed source and chopper.The time dependence of single energy groups and, hence, thetime moments were derived from these spectra. The measurementswere performed in a similar way as described in L6 - a7. Toimprove the time resolution beyond the chopper half width =50 s, the time delay between injection of the fast neutronburst and the chopper cut-off time was varied in steps of5 - lO/usec.

- 4: -

AED-Conf.-71-100-14

The experimental arrangement is indicated schematically inFig. 3. It is similar to the facility described by ReichardtL-8_/. The 14 MeV source neutrons (pulse width 2/usec) were mo­derated by a layer of 10 cm lead and 9 cm graphite in order tosoften the incident spectrum. The surface extraction area of thethermal neutrons was about 1.2 cm2 at half height of the cylin­der at a distance of 1.5 cm from the chopper disko

The moderators investigated were1) liquid hydrogen with 98% para-H2 (98%PLH2) at T = 20.4oK

2) liquid hydrogen with 77% para H2 (77%PLH2) at T = 20.40K

3) solid methane T = 770K.

The measured quantities were stationary spectra and thefirst two time moments. Time moments require integration of thetime dependent neutron flux to t ~ oo , This can be done if thereis an exponential decay and the decay constant is known. In para­hydrogen, no asymptotic decay could be observed. This resultedin uncertainties of the variances of up to 30%. A similar diffi­culty exists for all moderators in the transient region.

A more detailed description of the measurements and cal­culations will be published.

4. DISCUSSION OF THE RESULTSFig. 2 shows measured leakage spectra of cold moderators

compared w1th water at room temperature. The geometricalarrangement is indicated schematically in Fig. 3. For para­hydrogen, the di~er of the container used gives nearly theoptimum leakage intensity. For water at room temperature, thecontainer dimensions are much too big. Therefore, a waterspectrum is included in the same figure calculated for optimumgeometry as indicated in Fig. 6. This spectrum is normalizedto give the flux relative to parahydrogen which would exist inthe boester.

The calculated and the measured variances for differentmoderators are shown in Fig. 4. Measurements and two-dimensionalTHERMOS calculations are in good agreement. In the slowing downregion they show the well-known Ijv-behaviour L-9_/. For ener­gies below 13 meV, where the scattering cross section for para­hydrogen 1s very small, the mean free path becomes comparableto the moderator dimensions. Thus, time-of-flight effects leadto a I/V behaviour of the variance. This means that in time-of­flight experiments energy resolution below 13 meV is indepen­dent of energy. The other moderato~show an energy independentvariance in the thermal region which is nearIy equal to thethermal neutron lifetime.

Aseries of two-dimensional multigroup SN-calculationsin booster geometry to find the optimum moderator dimensionswere performed. For a cylindrical parahydrogen moderator of30 cm height, as proposed for the booster, the perpendicularleakage flux ~ for 10 meV neutrons and 1 W booster power isshown in Fig. 5 as a function of the radius of the moderator.

- 5 -_ .._- ---

AED-Conf.-71-100-14

The figure of merit

FOM = _---LIl1""":-_ ":"

(M 1) 2+ t)'2

with 1 M = 17.5 sec and~ for 10 meV neutrons has its maximumvalue at 6 cm radius. Similar optimization studies for liquidwater moderators of 30 cm height led to a heterogeneouslypoisoned arrangement. A 4 x 12 cm slab, with the large sideclose to the core surface, is subdivided into three 4 x 4 cmcells by two cadmium sleeves. In Fig. 6, the FOM for thiswater moderator is compared with the FOM of an optimized para­hydrogen moderator as a function of energy.

For energies below 15 meV, parahydrogen is superior towater by about a factor of 3.

An improvement seems to be possible by a more skillfalparahydrogen arrangement. Future experimental and theoreticalwork will deal with this question.

6

AED-Conf.-71-100·-14

References

Fast Burst Reactors, Albuquerque Conf. 690102,Jan. 28 - 30, 1969

W.L. Wittemore, A.W. McReynolds, G.A. 2503

A.E. Profio et al., Pulsed Neutron Research,lAEA-Vienna, Vol.I, p.123, 1965

C. Günther, W. Kinnebrock, private communication

J. Young, J. Koppel, Phys. Rev. 135 A, 603, 1964

E. Barnard et al., Nucl. Science and Engineering,17, 513, 1963

R.G. Kryter et al., Pulsed Neutron Research,IAEA-Vienna, Vol.I, p. 465, 1965

W. Reichardt, External Report KFK-INR-4 67-11, 1967

K.H. Beckurts, K. Wirtz, Neutron Physics,Springer 1964

- 7 -

AED-Conf.-71-100-14

- 8 -

AED-Conf.-71-100-14

o., ., y 98% PLH2 T=20,4 K

o0.0 77% PLH2 T= 20,4 K+ + + SOLID METHANE T=77° Kx x x H20 T=2930 K

THERMOS - 77% PLH2 --2-DIM CALCULATION 98% PLH2 -

5 2-DIM ~~I~~~C~:i~2~:§~TRY ENERGV meV2 I i I I I I i I I I

lÖ1 2 5 10° 101 102 103

FIG.2 ME'ASURED NEUTRON LEAKAGE SPECTRA I NORMAlIZED TO THE

SAME FLUX OF FAST INCIDENT NEUTRONS

CADMIUM LINED

MODERATOR

R:7.4cm

H:16 cm

CRYOSTAT

CHOPPER (200 r. p.s.)

4C - PRECOLLIMATOR1.1x1.1 cmxl0 cm

DIVERGENCE =3.30

FLIGHT-PATHL:3.1 m

"MeV

50x30x9 cm GRAPHITE

SOx 30x10cm LEAD

T(d~CEFIG.3 EXPERIMENTAL ARRANGEMENT

- 9 -

AED-Conf.-71-100-14

101 CALCULATION

0000 98% PLH25 +++ 77°'0 PLH2

vvv SOLID METHANE

522'---"'---........----I'--_J......_..L..-_---L_~ __---I:..----'

0.5 102ENERGY meV

FIG.4 VARIANCE OF THE THERMAL EMISSION TIME DISTRIBUTION

x i ~~ CI/l....J

'715 E 75UJ ~~ t"-o ~~ ':"""« - It)

UJ ~ ....52...J Ee::: 52 ~

« ..... 0

510 ..e- u, 50u....0zUJ0-e:::UJ0-

5 25

01oFIG.5

RADIUS [ern] I10

5 10 15 20CALCULATED FWX AND FOM FOR 10 meV NEUTRONS INLIQUID· PARAHYDROGEN

- 10-

AED-Conf.-71-100-14

WAIEB..: ij$=uCd-sleeves12x4 cm

"'_ . O,2cmHEIGHT H=30cm

4cm

MODERATORGEOMETRY:

PARAHYDROGEN : ~RADIUS R=6cmHEIGHT H=30 cm

1.0

0.5

10.0r--::=:---r------r--r---r----r----,--,.----,----,

Fig. 6 Comparison of the FOM for parahydrogen and water.The optimum geometry for both moderators is sche­matically indicated.

- - - --

- 11 -

-~- -~- -- -- - - --- -- - ---- --~--~~ --------~------~--- - -~ -----~- -~-~---------------~------ - ---~ ---~-- - - ~~ - ~- ----- - ~--- -~------~-~----_____i